Language-Guided Traffic Simulation via Scene-Level Diffusion
This addresses the need for easier-to-use traffic simulation tools for practitioners in autonomous vehicle development, representing a novel integration of language guidance into traffic modeling.
The paper tackles the problem of making traffic simulation more accessible and controllable for autonomous vehicle development by introducing CTG++, a scene-level conditional diffusion model guided by language instructions, which generates realistic and query-compliant traffic simulations as demonstrated in evaluations.
Realistic and controllable traffic simulation is a core capability that is necessary to accelerate autonomous vehicle (AV) development. However, current approaches for controlling learning-based traffic models require significant domain expertise and are difficult for practitioners to use. To remedy this, we present CTG++, a scene-level conditional diffusion model that can be guided by language instructions. Developing this requires tackling two challenges: the need for a realistic and controllable traffic model backbone, and an effective method to interface with a traffic model using language. To address these challenges, we first propose a scene-level diffusion model equipped with a spatio-temporal transformer backbone, which generates realistic and controllable traffic. We then harness a large language model (LLM) to convert a user's query into a loss function, guiding the diffusion model towards query-compliant generation. Through comprehensive evaluation, we demonstrate the effectiveness of our proposed method in generating realistic, query-compliant traffic simulations.